
RS
Robert Sutherland, Stefan Goetze, Jon Barker
· 1 min read
ResearcharXiv cs.CL
Challenges of Multi-Speaker Extraction for Real Conversational Speech Enhancement
arXiv:2609.25948v1 Announce Type: cross
Abstract: Target-speaker and multi-speaker extraction are techniques for extracting speech from a desired speaker or desired speakers in the presence of other speakers and/or noise. Neural network approaches for this task are often trained and evaluated using simulated datasets, with balanced amounts of target speech and speaker enrolment samples which closely match the target speech. However, in real multi-party conversations, participants are often silent for more time than they are speaking, and their enrolment speech samples can differ substantially from the target speech in the conversation. These factors can impact the training and evaluation of these techniques on recordings of real conversations. This work proposes a new loss function, which helps mitigate the effect of excess silence in training, improving STOI from 0.55 to 0.60, and frequency-weighted segmental SNR from 4.35 to 5.12. Additionally, the impact of the mismatch between the enrolment speech and target speech is explored.
Original source
This story was published by arXiv cs.CL and written by Robert Sutherland, Stefan Goetze, Jon Barker. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on arxiv.org


